結構物強震反應下之系統識別方法比較

Comparison on Structure Identification Method for Strong Earthquake Response Data

葉乃睿(Nai-Ruey Yeh)

指導教授 ： 羅俊雄

國立臺灣大學/工學院/土木工程學研究所/碩士(2016年)

https://doi.org/10.6342/NTU201600852

摘要

研究的目的在於分析結構物地震反應訊號，經由系統識別的方法來識別結構物在地震時的動態特性以及其時變性，透過了解結構物動態特性的變化，可以判斷結構物是否有破壞產生，這就是結構健康監測最主要的目的（Structure Health Monitoring）。
本研究主要是使用（1）子空間識別法、（2）遞迴子空間識別法、（3）遞迴ARX model（4）移動型窗函數子空間識別法來對結構物進行系統識別，透過子空間識別法來瞭解結構物在長期監測下棋動態特性有無發生變化，而遞迴子空間識別法、遞迴ARX model、移動型窗函數子空間識別法主要是用來即時監測結構物在地震過程中動態特性的時變性，但這些方法都各有其優缺點，而本研究的重點就是比較各方法之間的優缺點並利用其優點來清楚並正確的識別結構物的動態特性，本文所使用的地震資料均為中央氣象局的強震監測資料。

關鍵字

子空間識別法 ； 遞迴子空間識別法 ； 遞迴ARX model ； 移動型窗函數子空間識別法 ； 結構健康監測

並列摘要

This study is mainly to analyze the earthquake response data, identify the structure dynamic characteristics by system identification methods.We can judge whether the structure dynamic characteristics is changed by the earthquake or not, and this is the main objective of the Structural Health Monitoring.
This study apply four methods on the earthquake response data: (1)Subspace Identification, (2)Recursive Subspace Identification, (3)Recursive ARX model, and (4)Moving window Subspace Identification.We can realize the changes of structure system parameter by the long‐term seismic response monitoring of structures using subspace identification. It can help us to know the structure dynamic behavior in the past. And the other three methods can help us to realize the change of system parameters. If there are some damage in this structure,the system parameters will change, that include (1) natural frequency, (2) damping ratio, and(3)mode shape. However , these methods are all has its advantage and disadvantage,and this study also focusing on the comparison on these four methods. All the data we use in this study are from the Central Weather Bureau Structure Strong Earthquake Monitoring System.

並列關鍵字

Structure Health Monitoring ； Subspace Identification ； Recursive Subspace Identification ； Recursive ARX model ； Moving window Subspace Identification

參考文獻

[25]翁健煌.(2010).子空間識別法於系統識別及結構損壞診斷之應用. 臺灣大學土木工程學研究所學位論文, 1-183.

[2]Dickinson, B., Kailath, T. and Morf, M. (1974). Canonical Matrix fraction and state space descriptions for deterministic and stochastic linear systems, IEEE Transactions on Automatic Control, AC-19 656-667.

[3]Gram, J. P. (1883). Uber die Entwickelung reeler Functionen in Reihen mittelst der Methode der kleinsten Quadrate,Journal fr die reine und angewandte Mathematik,94, 41-73.

[6]Gustafsson, T. (1997). Recursive system identification using instrumental variable subspace tracking, Proc. of the 11th IFAC Symposium on System Identification, Fukuoka, Japan.

[9]Lovera, M., Gustafsson, T., and Verhaegen, M. (2000). Recursive subspace identification of linear and non-linear Wiener state-space models, Automatica, 36, 1639-1650.